OpenAI response brief
Read-only OpenAI Responses API workflow. It counts input tokens, creates one explicit non-streaming response through recued-core.openai.response.create with store=false by default, and summarizes status, output items, and usage. It never enables streaming, hosted tools, custom function tools, MCP tools, file uploads/downloads, response deletion, background cancellation, vector stores, realtime, or arbitrary OpenAI routes.
How it works
Inspect the data fetches, transforms, gates, and output this recipe runs.
Process (7 steps)
input_tokens
?
response
?
output_items
default
Apply default
output_count
count
Count items in output items
summarize
?
output_table
to_table
Format results as a data table
card
to_summary
Format results as a summary card
Settings
Configurable at install. Defaults shown — change them anytime in Recued.
focus
setting
=
Summarize the OpenAI response, status, output item types, token usage, and practical prompt/model follow-ups. Do not invent facts outside the OpenAI API response.
model
setting
=
gpt-5.5
openai
setting
=
prompt
setting
=
reasoning
setting
=
[object Object]
max length
setting
=
900
temperature
setting
=
0.2
text config
setting
=
[object Object]
instructions
setting
=
Return a concise, directly usable answer. Call out uncertainty and avoid inventing facts not supported by the supplied input.
service tier
setting
=
auto
store response
setting
=
false
max output tokens
setting
=
900
Trust & control
What installing this recipe would let it do. Recued grants these permissions at install — review them there before approving.